by: John C. Purvis
Evaporation is an important, but often-overlooked, climate element. Evaporation affects both plant and animal life and is a major factor in man's comfort and well-being. In spite of its importance, evaporation is routinely measured at only a few meteorological stations in South Carolina.(1)
The usual way of measuring evaporation is through the use of a type of atmometer, an evaporation pan (2). The National Weather Service's Class-A Evaporation Pan is a cylindrical container fabricated of monel metal with a depth of ten inches and a diameter of forty-eight inches. The pan is leveled at a site that is usually well-sodded and free from obstructions. The pan is filled with water to a depth of about eight inches, and daily measurements are made of the water level. When the water level drops to seven inches, the pan is refilled. Daily measurements are corrected as necessary for rainfall, and in some cases water temperature. The resulting daily water loss is the pan evaporation for the period considered, usually 24 hours.
Pan evaporation is usually greater than the actual evaporation from nearby land surfaces. A widely accepted coefficient of pan evaporation to the actual evaporation is approximately 0.70(3). For example, if pan evaporation on a typical day were 0.20 inches and the coefficient of pan evaporation were 0.70, the true evaporation would be approximately 0.14 inches (0.20 x 0.7). One reason that the pan evaporation measurements average more than evaporation from nearby land surfaces is that the evaporation pan always presents a freely evaporating surface to the atmosphere. Another factor is that the sensible-heat transfer through evaporation pans can be appreciable and tends to increase the pan evaporation.
Prompted by the need for evaporation data, the Southeast Regional Climate (SERCC), in cooperation with the National Climatic Data (NCDC), initiated an evaporation study of pan evaporation data collected in the Southeast. The purpose of this study is to increase the reliability of this data and to enhance usage in meeting the evaporation needs of agriculture, industry, state water planning, and related research demands.
The first work is limited to pan evaporation data collected in South Carolina and nearby stations located in North Carolina and Georgia. Once operating quality control and estimation procedures have been finalized and techniques developed for meeting these objectives, the study will be expanded to include pan evaporation data available across the Southeast.
The first location selected for this study was Clemson University's Edisto Experiment Station. This evaporation station is located in southern South Carolina approximately three miles west of Blackville, South Carolina. Edisto Experimental Station is in a rural area and is representative of a large progressive agricultural area.
An examination of pan evaporation records from the Edisto Experiment Station revealed some missing data, mostly of a one to two day event occurrence. It was decided to use a method devised by Thornthwaite and Mather(4) using temperature to calculate unadjusted potential evapotranspiration. This is basically an empirical measurement of the evaporative power of the air. Other more accurate methods using solar radiation were investigated but not used due to the lack of solar radiation measurements. Thornthwaite's technique may not be the most accurate method, but it does surprisingly well in the Southeast(5). It operates on the assumption that mean temperature and day length represent the most important criteria which influence potential evaporation and that all other factors tend to average out in most cases over an extended period. Based on the above assumption (6), the mean temperature (Figure 1) and the day length in units of 12 hours (Figure 2) were calculated for each occurrence of missed pan evaporation data. Using this information the unadjusted potential evapotranspiration (PET) was computed for each date of missing data.
Since evaporation pans tend to increase the actual pan evaporation (7), it is necessary to inflate the unadjusted potential evaporation data computed by the Thornthwaite method, before substituting these data for the missing pan readings. Kohler, Nordenson, and Bader computed evaporation pan coefficients for the entire nation. Kohler's values were used to adjust the potential evapotranspiration to reflect pan evaporation (Figure #3). These adjusted values were then substituted for the missed entries.
1. ANALYSIS OF DATA:
After all short periods of missing data were computed, pan evaporation data for the various sites in or near South Carolina were grouped, as available, for each decade beginning with 1950. Since the period of study ended with the 1992 records, the decade beginning with 1990 only included 1990-1992 information. Comments concerning this grouping are as follows:
A. ATHENS COLLEGE OF AGRICULTURE, GEORGIA (See Table #1)
(1) Number: 09043202
(2) Lat., Long.: 33 55N 083 22W
(3) Elevation: 689ft MSL
(4) Time of Obsn: unknown
(5) Significant Moves: Moved to University of Georgia Plant Science Farm in 1971.
B. UNIVERSITY OF GEORGIA PLANT SCIENCE FARM (See Table #2)
(1) Number 09895002
(2) Lat., Long.: 33 52N 083 22W
(3) Elevation: 840ft MSL
(4) Time of Obsn: 08
(5) The University of Georgia Plant Science Farm is sevaral miles southwest of the Athens College of Agriculture observation site. The combined observation record from these two sites cover the period 1953-1992. It is difficult to isolate a definite trend although the measurements from the University of Georgia Science Farm after 1971 are slightly higher than those from the Athens College Station. It is interesting to note that these locations are a higher elevation than the Clark Hill site. This should contribute to a lower evaporation rate at these two stations than that of Clark Hill. In sharp contrast to this expectation, however, the reverse is true. The readings from the Georgia stations are not lower but considerably higher than Clark Hill's.
B. CATALOOCHEE, N.C. (See Table #3)
(1) Number 31156401
(2) Lat., Long.: 35 37N, 83 6W
(3) Elevation: 2620ft MSL
(4) Significant Moves: Before 1972 the evaporation pan was located at Lat. 35 38N, Long. 083 05W. Although located outside South Carolina pan evaporation measurements from this station give some indication of what might be expected in the higher elevations of the mountains of South Carolina. Readings during the colder period of the year, December-March are not available. No definite trends were noted in the Cataloochee records.
C. CHAPEL HILL, N.C. (See Table #4)
(1) Number 31167703
(2) Lat., Long.: 35 55N, 79 06W
(3) Elevation: 500ft MSL
(4) Time of Obsn.: 08
(5) Significant Moves: none
Due to the breaks in the Chapel Hill data, it is impossible to define any specific trend. If South Carolina isolines were extended into North Carolina, the average decadal pan evaporation values are approximately what would be expected.
D. CHARLESTON, S.C. (CHARLESTON AIRPORT, See Table #5)
(1) Number 38154407
(2) Lat., Long.: SC 32 54N, 80 02W
(3) Elevation: 40ft MSL
(4) Time of Obsn.: Mid
(5) Significant Moves: Evaporation pan was relocated on the other side of the Airport in 1982. Changes in elevation and exposure were minor.
(6) Remarks: Charleston's pan evaporation records were averaged for each decade beginning 1960. These decadal annual averages showed an upward trend with January 1970-79 having the lowest values.
E. CLARK HILL, S.C. (See Table #6)
(1) Number: 38172605
(2) Lat., Long., 33 40N, 82 11W
(3) Elevation 380ft MSL
(4) Time of Obsn.: 08
(5) Significant Moves: none
The pan evaporation records for Clark Hill differ considerable from those of surrounding stations. This anomaly merits further study. The Clark Hill pan evaporation site is downwind from Thurmond Lake, a large hydro-electric facility on the Savannah River. It appears that this site is not representative of a large portion of the Savannah Valley. Decadal annual pan evaporation averages for Clark Hill decreased through 1979 but increased significantly during the 1980-89 period. The period 1990-92, however, showed a decrease in all monthly averages except January and February which revealed an upward trend.
F. CLEMSON, S.C. (see Table#7)
(1) Number 38177002
(2) Lat., Long.: 34 41N 082 49W
(3) Elevation: 819ft MSL
(4) Time of Obsn.: 08*
Observation time was 1700 from beginning of observations until July 1963, and from October 1965 to April 1973.
(5) Significant moves: Before August 1973 the evaporation pan was located at Lat. 34 40N, Long. 082 50W.
Clemson's pan evaporation data were summarized in decadal form from 1950 to 1992. Annual averages increased through 1979 but have shown a downward trend beginning with the 1980-89 decade. The average monthly decadal values for January and February, however, do not reflect the same trend, and have fallen from 1950 through the decade ending in 1989, but increased during the 1990-92 period.
G. EDISTO EXPERIMENTAL STATION (Blackville 3W, See Table #8)
(1) Number: 38076407
(2) Lat., Long.: 33 22N, 81 19W
(3) Elevation 324ft MSL
(4) Time of Obsn.: 17
(5) Significant Moves: The Edisto Experiment Station located three miles west of Blackville began pan evaporation measurements in 1963. The average annual pan evaporation during the periods 1963-1969 and 1970-79 increased but showed a slight decrease during the 1980-89 decade. This decrease was more significant during the summers of the 1990-92 period.
H. FLORENCE 8NE (See Table #9)
(1) Number: 38311104
(2) Lat., Long.: 34 18N, 079 44W*
(3) Elevation 120 ft MSL
(4) Time of Obsn.: 08
(5) Significant Moves: * Before July 1983 the pan evaporation station was located at Lat. 34 13N, and Long. 079 46W. The station moved 6.5 miles from 2 miles north of Florence to 8 miles northeast of Florence. There is no significant change in soil type or topography between the two stations.
Florence's period of observations are relatively short but correspond nicely with nearby pan evaporation observations. There was a slight increase in annual averages during the period of record, although the average monthly changes were mixed with a few months reporting lower values.
I. HOFMAN FOREST, N.C. (See Table 10)
(1) Number 31414406
(2) Lat., Long., 34 50N 77 18W
(3) Elevation 44ft MSL
(4) Time of Obsn: 19
(5) Significant Moves: none
This station is located in extreme eastern North Carolina. Unfortunately, the records are not long enough to establish any trend.
J. LUMBERTON 3SE, N.C. (See Table #11)
(1) Number 31517706
(2) Lat., Long.: 34 37N 078 59W
(3) Elevation 112ft MSL
(4) Time of Obsn. 08
(5) Significant Moves: During the period 1958-1960 the station was located at Lat. 34 42N, Long. 079 00W. Lumberton's relatively large pan evaporation measurements are more than those made at nearby stations. This is not unexpected, however, due to the location of Lumberton to the Sandhill area of North Carolina. There has been no significant change in Lumberton's annual pan evaporation readings during the period of record.
K. SANDHILL RESEARCH STATION-ELGIN (See Table #12)
(1) Number 38766606
(2) Lat., Long.: 34 08N 080 52W
(3) Elevation 440ft MSL
(4) Time of Obsn.: 08
(5) Significant Moves: none
The pan evaporation records for the Sandhill Research Station were consistently higher than pan evaporation measurements in the upper coastal plain or the lower piedmont. These higher evaporation measurements at the Sandhill Station result from the effect of the sandy terrain. The sandy soils dry out more rapidly than clay or loamy soils and during the summer heat more rapidly with accompanying lower relative humidities and higher evaporation rates. In reviewing the Sandhill records, no definite trend was noted. The average annual decadal pan evaporation decreased in the 1970s, increased in the 1980s but decreased thereafter.
L. SAVANAH WEATHER SERVICE OFFICE AIRPORT (See Table #13)
(1) Number 09784709
(2) Lat., Long.: 32 08N 081 12W
(3) Elevation 046ft MSL
(4) Time of Obsn.: 07
(5) Significant Moves: unk
Observations for the 1965-69 period were slightly larger than the decadal average for 1970-79. There has been, however, a slight upward trend since 1980. This upward trend agrees quite well with the measurements made at Charleston.
M. UNION 7SW (SOUTHEAST FOREST EXPERIMENT STATION, See Table #14)
(1) Number 38878602
(2) Lat., Long.: 34 38N 81 40W
(3) Elevation 500ft MSL
(4) Time of Obsn.: 07
(5) Significant Moves: Station was relocated in 1964 to a new site with different exposure. Evaporation observations were discontinued at the time of relocation. The relatively short period of pan evaporation measurements from Union conform quite well with readings from surrounding stations. The period is too short, however, to confirm any definite trend.
2. VARIATION OF PAN EVAPORATION ACROSS SOUTH CAROLINA
There is considerable variation in the average monthly and annual pan evaporation measurements across South Carolina (Fig 4-8). The largest annual pan evaporation readings during the period of record are in the extreme southeast part of South Carolina where annual totals exceed 60 inches. The average annual 60 inch isoline enters the State south of Myrtle Beach and proceeds southwest into Georgia about 40 miles northwest of Savannah. Pan evaporation elsewhere over the Coastal Plains is between 50 and 60 inches per year. A second area of pan evaporation readings exceeding 60 inches per year is in the Sandhills or the narrow northeast-southwest belt that extends through the central part of the state and corresponds to the location of the Fall Line separating the Coastal Plain from the Piedmont. Moving northwest across the Piedmont, pan evaporation measurements decrease with annual totals of around 40 inches in the higher mountains of northwest South Carolina. The only significant anomaly to the above are the Clark Hill ovservations which are lower than nearby areas. The Clark Hill anomaly merits further study. The Clark Hill evaporation pan is immediately downwind from Thurmond Lake. It is unlikely that this observation station is representative of a large portion of the state.
This pan evaporation data has been summarized for each decade. The Charleston data shows an increase each decade with the greatest increase occurring in the three year period 1990-92. A similar increase was noted at Savannah, Ga. This latter site, although located in Georgia, is the nearest coastal pan evaporation observation site to Charleston.
Pan evaporation observations in northwest South Carolina do not confirm the upward trend noted in southern South Carolina. Clark Hill pan evaporation values have decreased in recent years. Clemson's readings increased through 1979 but have fallen since then.